AI for IBD v2.0
Research type
Research Study
Full title
Real-time artificial intelligence IBD-Scoring solution for Inflammatory Bowel Disease
IRAS ID
299417
Contact name
Marietta Iacucci
Contact email
Sponsor organisation
University of Birmingham
Duration of Study in the UK
2 years, 6 months, 4 days
Research summary
Inflammatory Bowel Disease (IBD) is a chronic, relapsing disease and includes two disorders. Ulcerative Colitis (UC) affects the colon, whereas Crohn’s Disease (CD) can involve any part of the gastrointenstinal tract. Patients are monitored but the assessment of disease activity is challenging. Assessment involves endoscopic assessment and histological assessment. Disease activity is scored using a range of scores, but current scores rely on the subjective assessment of the observer. Artificial intelligence clinical decision support solutions (AI-CDSS) carry the potential to mitigate operator-dependent limitations in endoscopy and histology and provide an easily accessible tool that delivers an assessment of mucosal healing/disease activity that correlates with endoscopic and histological remission.
Therefore, this single centre observational study will collect data to contribute to the development of robust AI models. This tool will do real-time IBD disease assessment during endoscopy to aid the pathologist and to help predict future disease course and reponse to treatment.
Potential participants with a diagnosis of IBD who are scheduled to receive a colonoscopy will be invited to join the study. During the endoscopy, video and still images will be captured. Standard care biopsies will be taken, and digital images taken from biopsy slides. Videos and images, together with related scores and basic demographic and disease data will be sent to Satisfai Inc to be used as training and validation datasets.
The primary outcome will be the diagnostic performance of the AI-CDSS platform for assessing IBD disease activity on endoscopic and histologic bases.
Participants will attend endoscopy as required for management of their disease and no additional study related visits are required. All data will be pseudoanonymised using a unique study identifier to allow linking between datasets. The key to the study identifier will remain at the study site.REC name
London - Westminster Research Ethics Committee
REC reference
21/PR/1063
Date of REC Opinion
27 Aug 2021
REC opinion
Further Information Favourable Opinion